Background/Purpose: Bone marrow lesions (BML) and synovitis
are MRI features that have been associated with symptoms and disease progression
in patients with osteoarthritis. The Hip Inflammation MRI Scoring System
(HIMRISS) has been developed to specifically score these active lesions in hip
OA. We aimed to conduct validation according to the OMERACT framework of a
novel web-based overlay integrated to facilitate sectoral
scoring of these lesions in patients receiving intra-articular hyaluronate.

Methods:   We conducted MRI at
baseline in patients receiving intra-articular hyaluronate
for symptomatic hip OA. A circular overlay integrated with a web-based DICOM
viewer was designed for use with DICOM MR images of the hip for evaluation
according to the HIMRISS method. Using a mouse, the overlay can be positioned
over the largest diameter of the femoral head and resized to ensure alignment
with the subchondral bone. BML are assessed on
coronal fluid-sensitive MR scans using an overlay comprising a circle divided
into octants. The presence/absence of BML in each of the octants is scored
dichotomously directly on the image display in 5 consecutive coronal slices,
simply by mouse-clicking on any octants containing
BML, thereby dispensing with scoring sheets. The overlay is also used to score BML
in the acetabulum, assessed in 3 sectors. Synovitis
is scored in 2 locations according to a grading scheme (0-2) that reflects
thickness of synovitis-effusion.

Intra-class correlation coefficients (ICC) were
used to assess reliability between 4 readers without any formal calibration.
Correlation analysis, univariate and multivariate
analyses, adjusted for age, sex, symptom duration, BMI, and baseline WOMAC
pain, were used to assess the predictive capacity of baseline BML and synovitis scores for change in WOMAC pain from baseline to
12 weeks.

Results: The cohort included 60 patients, 28 (46.7%) males, mean (SD) age
of 63.0 (11.2) years, mean (SD) disease duration of 31.4 (43.4) months,
baseline mean (SD) WOMAC pain of 49.7 (16.3), mean (SD) change in WOMAC of
-17.9 (19.2), mean (SD) femoral BML of 10.6 (13.9) (0-65 is maximum possible
range), mean (SD) acetabular BML of 7.0(5.5) (0-35 is
maximum possible range), mean (SD) effusion of 13.3 (8.1) (0-30 is maximum
possible range).

Inter-observer ICC [95% CI] was
0.83[0.78-0.87] for femoral BML, 0.64 [0.52-0.74] for acetabular
BML, 0.78 [0.65-0.86] for synovitis-effusion, and
0.86 [0.81-0.89] for total HIMRISS score. Significant correlation was evident
between baseline femoral BML score and change in WOMAC pain (r =0.44, p=0.001).
 In univariate
analysis, baseline femoral BML (β=0.58,
p=0.0009) was significantly associated with change in WOMAC
pain, an association which remained present in multivariate analysis (β=0.51,
p=0.01). 21/55 subjects (38.2%) had a 50% reduction in WOMAC pain and 39/55
(70.9%) a 20% reduction in WOMAC pain. Baseline femoral BML predicted a WOMAC
20% response (OR [95%CI]: 0.95 [0.91-0.99], p=0.018) in multivariate logistic
regression.

Conclusion:
The web-based adaptation of HIMRISS facilitates reliable scoring of active
lesions in hip OA and is relevant to patient symptoms and response to
treatment.